Access to benefits from forest commons in the Western Himalayas

Ecological Economics 71 (2011) 202–210

Contents lists available at SciVerse ScienceDirect

Ecological Economics journal homepage: www.elsevier.com/locate/ecolecon

Analysis

Access to benefits from forest commons in the Western Himalayas Sirisha C. Naidu ⁎ Department of Economics, Wright State University, 3640 Colonel Glenn Highway, Dayton, OH, 45435 USA

a r t i c l e

i n f o

Article history: Received 13 October 2010 Received in revised form 28 August 2011 Accepted 7 September 2011 Available online 8 October 2011 Keywords: Resource access Community forestry Equity Resource dependence Rural livelihoods India

a b s t r a c t Little statistical evidence exists on the effects of forest management regimes and wealth on forest access rates in South Asia. To determine the magnitude and significance of these effects, this paper analyzes a dataset of communities from Himachal Pradesh, India, with a fractional logit model. The investigation considers three specific forest management regimes including a regime under complete state control, traditional community regime and a co-management regime known as Joint Forest Management. Communities with higher incidence of land poverty have lower forest access rates for grazing and fodder extraction, whereas communities with a higher incidence of land-rich households have higher forest access rates for fodder extraction. Forest access rates for fuelwood collection are lower under traditional and co-management regimes. However, the interaction between land-poverty and co-management regime increases forest access rates for fodder collection and livestock grazing. © 2011 Elsevier B.V. All rights reserved.

1. Introduction The bulk of empirical research on the issue of forest access has focused its attention on uncovering how socioeconomic characteristics of users influence forest access (Adhikari, 2005; Adhikari B. et al., 2004; Beck, 1994; Beck and Ghosh, 2000; Beck and Nesmith, 2001; Cavendish, 2000; Coulibaly-Lingani et al., 2009; Jodha, 1986, 1995; Kamanga et al., 2009; Khan and Khan, 2009; Mamo et al., 2007; Narain et al., 2008a,b). Although this focus on internal social structure is essential, most studies are relatively silent about the effect of forest management regimes on access. The issue is crucial: management regimes determine access and distribution of benefits from resources (see Larson, 2008; Larson et al., 2010; Ribot and Larson, 2005). In South Asia, management regimes are particularly of interest due to a rise in the popularity of community forestry initiatives. Indian forest policies, for instance, have adopted community forestry in a remarkable turnaround from policies with roots in colonial “scientific” management. Touted as a win-win approach for rural livelihoods and forest conservation, structured community forestry initiatives in India have received the financial blessings of international aid, development and conservation agencies. According to one estimate, the adoption of community forestry in the country has resulted in its coverage of 27% of the national forest area across 85,000 village committees (World Bank, 2005). Advocates of the community forestry have pitched this policy to developmentalists as an approach that potentially increases access

⁎ Tel.: + 1 937 775 2960; fax: + 1 937 775 2441. E-mail address: [email protected]. 0921-8009/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.ecolecon.2011.09.007

to forest benefits and hence improves rural wellbeing (World Bank, 2005). It is argued that this policy will foster equitable and inclusive outcomes on account of devolution of power to local communities (e.g., Bromley, 1992; Ostrom, 1990; Wade, 1988). In light of research suggesting that the much vaunted community forestry programs in South Asia are associated with unequal forest access (Adhikari, 2005; Adhikari B. et al., 2004; Agarwal, 2001, 2007, 2010; Beck and Ghosh, 2000; Iversen et al., 2006; Thoms, 2008; Wilshusen, 2009), there are doubts about the desirability of this policy. Yet research into the level of access across different forest management regimes has received sparse attention (but see Adhikari M. et al., 2004). A central objective of this paper is therefore to examine the effects of forest management regimes on access to forest benefits in forest communities in the Indian Himalayas. Although forests in India are state property, co-management initiatives undertaken by the state have not fully replaced state management regimes; these two regimes, moreover, exist alongside traditional community forest management regimes. Utilizing the statistical method of generalized linear models (GLM), the paper tests whether forest access differs across state, co-management and traditional forest management regimes. A second objective is to investigate the effects of wealth on forest access and specifically to ascertain the impact of community forestry on forest access to the land-poor households. The paper aims to contribute to the discussion on the desirability of community forestry initiatives. The paper is organized as follows. The next section describes the data collection and survey methods utilized. Section three discusses the concepts of forest access and forest regimes and describes the sample. Econometric results are reported in section four. The paper concludes with a discussion of the findings and policy implications.

S.C. Naidu / Ecological Economics 71 (2011) 202–210

203

Fig. 1. Map of Himachal Pradesh. Source: Forest Survey of India, 2009.

2. Study Site, Data Collection and Survey Methods Data were collected in 2004 during fieldwork in Himachal Pradesh, a state situated in the western Himalayan region of India. Despite being a relatively small state, with a total geographical area of 55673 km2, there is significant variability in altitudes (i.e., 350–6795 meters above sea level), climate and geology, contributing to an ecologically diverse environment (GoI, 2001). According to the 2001 Census, 91% of the approximately six million population is rural (GoI, 2001). Eighty-seven percent of the rural population is dependent of forests for a variety of daily requirements (Gouri et al., 2004); this underscores the importance of forests for rural livelihoods. As part of a larger study aimed at investigating collective management and forest use, four sub-watershed regions in the districts were purposively selected on account of the variability in the region's forest management regimes. Usufruct rights in both districts lie with the community rather than the individual (see Chhatre, 2003; Morrison, 2001); thus households in a forest community1 have equal rights to forest extraction. This characteristic allows for an investigation of group or social outcomes, i.e., it allows the analysis of differences in the extent of forest use across forest communities with comparable forest rights. Thus, the forest community is the unit of observation in this paper (Fig. 1). 1 The community is defined as a group of households with rights to community forests (also see Agrawal and Gibson, 2001).

After consultation with local NGOs and Forest department officials in the selected areas, a sampling frame was constructed of forest communities with the following characteristics: a) their location was in the Middle Himalayan range (1000–2200 meters above mean sea level); b) communities were not engaged in commercial extraction of forest resources; c) communities were not engaged in explicit conflict with outside agents 2; and d) communities were situated within two kilometers of a forest. These conditions enabled control of geographical diversity and required the sampling frame to be consistent with the population of interest. A random sample of 56 communities was drawn from the sampling frame though there are missing data in not more than four cases. Data were collected from March 2004 to September 2004. During this time, the author made multiple trips to the selected communities to gain a better understanding of the local community and gain the trust of community members. The survey questionnaire was pretested during these trips and modified for efficiency. The formal data collection process used semi-structured group interviews.3 Each survey group

2 Mining is carried out in some forests. Mining contracts, however, have been awarded to outside agents with significant political influence and this fact is a source of conflict. Communities engaged in such conflict were excluded from the sampling frame. 3 The interviews were typically conducted on the day of village community meetings.

204

S.C. Naidu / Ecological Economics 71 (2011) 202–210

consisted of village community members. At least one official from the local administrative body (panchayat), or village forest committees, or local women's councils (mahila mandal) participated in the group interviews. These officials are typically residents of the village and the local institutions that they represent are often de facto local decisionmaking bodies over forest issues. Hence, these officials possess knowledge about social norms, and actual forest use, and management and importantly have access to aggregate data about the village community. All questions were posed to the group as a whole, were discussed among members of the entire group and answers recorded only after consensus was achieved.4 This mode of data collection reduced the amount of time required to gather data about characteristics of the forest community and was possible because of high levels of social interaction and interdependence among households in a community.5 Questions posed concerned the community in aggregate, i.e., about the participation in community institutions and access rates to forests in the community. This interdependence and the social interactions with members of the community prior to the formal interview provided little incentive to misrepresent information. The variables to be analyzed in this paper are shown in Table 1. 3. Forest Access and its Determinants: Concepts and Data Forest access is sometimes associated with property rights. It is thus assumed that the assignment of rights will lead to higher benefits. However, scholars have argued for a conceptual distinction between “rights”, “control” and “access”. Ribot and Peluso (2003) draw a distinction between property rights and access in order to argue that access “is about all possible means by which a person is able to benefit from things”; this includes paying attention to different institutions, and socioeconomic factors that allow or restrict flow of benefits. Property, on the other hand, deals with socially acknowledged legal or customary claims and is only one factor among many that allow resource users to benefit from the resource (Ribot and Peluso, 2003; Schlager and Ostrom, 1992). Advocates of decentralization and structured community-based natural resource management sometimes view community forestry initiatives as a change in property from the state to the local communities and as a pre-condition for acquiring benefits from forests. However, at least in the case of India, this has not been historically true (e.g., Gadgil and Guha, 1995; Guha, 1990). Despite the legal designation of forests as state property, forest dwellers in the Indian Himalayas have exerted individual or community efforts to attain and maintain control over access (Chhatre and Saberwal, 2006; Saberwal, 1999; Vasan, 2003, 2006). In the case of individual efforts, these may take the form of illegal extraction and incurring sanctions when caught, or cultivating personal relations (Ribot and Peluso, 2003) with Forest Department officials. Community efforts include setting up and maintaining forest management regimes. Forest management regimes “control”, direct and shape forest access irrespective of the rights that individuals or households claim (Larson et al., 2010; Meynen and Dornboos, 2005). Agarwal (2001; 2007; 2010), for instance, has devoted a considerable amount of effort in her research to highlight that women suffer from unequal access even if property rights exist. Hence efforts in this direction are significant. If higher social legitimacy is accorded to these traditional community management regimes, it could effectively transfer control of state-owned forest resource to the community (Rangan, 1997). 4 All questions pertained to the community as a whole, group interviewees were not asked for what might be considered private information about specific individuals or households. For instance, questions were asked about the number of households engaged in forest extraction, but not about the specific quantity of extraction. 5 Social interaction, even among members of different castes is high, especially on forest issues. Further women enjoy considerable autonomy and mobility and participate in meetings (Cranney, 2001 ; PROBE Team, 1999). Thus, women and lower caste community members were active participants in the group interview. In conducting the interview, the author also solicited information and opinions of lower caste and female members in case participation was passive.

Table 1 Variable names and descriptions. Proxies for access to forest benefits prop_ graz prop_ fodd prop_ fuel

Proportion of households grazing livestock in community forests Proportion of households extracting for fodder for livestock Proportion of households collecting fuelwood from community forests

Explanatory variables prop_lcaste prop_ poor

Proportion of lower caste households in the village community This variable measures the proportion of households unable to supply enough food for the entire year prop_landpoor Proportion of households that have landholdings below 1 acre prop_landrich Proportion of households that have landholdings above 2.5 acres prop_landcash Proportion of agricultural land in the community used for commercial purposes – indicates engagement with the agrarian output market shop_d Dummy variable that takes the value one if a fair price shop is located within 2 kilometers of the village shop_rich Interaction term between shop_d and prop_landrich traditional_d Dummy variable that takes the value one if traditional community management regime exists in the community and zero otherwise jfm_d Dummy variable for co-management regime set up by the Forest Department (in conjunction with NGOs and/or aid agencies) jfm_landpoor Interaction term between jfm_d and prop_landpoor

Similarly, Adhikari B. et al. (2004) and Adhikari (2005), in investigating the distribution of benefits under community forestry in Nepal, implicitly adopt the distinction between rights and ability to benefit, i.e., forest access. In the context of Himachal Pradesh and the study area in particular, rights to forests are comparable across all households in forest communities. Hence, this factor is held constant. However, even in the presence of equal forest rights, households and communities may differ in their access to forests because of wealth, forest management regimes, and other socioeconomic factors. In this paper, the discussion is restricted to an analysis of forest management regimes and wealth. This section briefly discusses the conceptual framework motivating the econometric specification and the data. 3.1. Access to Forest Benefits Forests in India were nationalized during British colonialism and have since been managed by a strong forest bureaucracy. While in most parts of the country, the Forest Department has legal control over forest management and access and has followed a policy of exclusion, sociopolitical realities in Himachal Pradesh allowed the maintenance of traditional usufruct rights of forest dwellers. These traditional rights, in both the districts of Mandi and Kangra, are vested with the community rather in individuals or households (Chhatre, 2003; Morrison, 2001). Therefore, all households in a community have usufruct rights to the community forest in order to graze livestock, and collect fodder, firewood, manure, bushes for fencing, timber for agricultural implements and non-timber forest products such as fruits, flowers, roots, bark, bamboo and honey (Chhatre, 2003). However, whether households have the wherewithal to benefit from this right (i.e., the ability to extract) and the absolute quantities extracted may vary due to management regimes and other socioeconomic factors (Ribot and Peluso, 2003; Schlager and Ostrom, 1992; also see Coulibaly-Lingani et al., 2009). Existing literature measures access to forest benefits in terms of absolute or relative income6 (e.g., Adhikari B. et al., 2004; Adhikari, 2005;

6 Literature on access and benefits from forests has spent significant effort to distinguish between absolute level of extraction and forest dependence (e.g., Coomes et al., 2004; Narain et al. (2008a,b) but the difference between the two is not addressed in this paper. Also note that I use the terms access, forest use, and forest benefits interchangeably.

S.C. Naidu / Ecological Economics 71 (2011) 202–210

Cavendish, 2000; Coomes et al., 2004; Coulibaly-Lingani et al., 2009; Khan and Khan, 2009; Mamo et al., 2007; Narain et al., 2008a,b; Reddy and Chakravarthy, 1999; however Jodha (1986) notes alternative measures). Notwithstanding measurement issues, which have benefited from advances in environmental valuation techniques (Cavendish, 2000; Vedeld et al., 2004, 2007), representation of the value of natural resources in market rubrics is fraught with problems. One concern is that the use value of forests (i.e., their ability to satisfy human needs) is not adequately incorporated into exchange value (Ackerman, 2008; Burkett, 2006; Vlachou, 2002), expressed either in market or in shadow prices. The objection is of particular relevance in the sampled communities where extraction primarily occurs outside the market, and socioeconomic relations are not solely governed by the logic of individual rationality. Further, aggregate or average forest income does not indicate the extent to which households have access to forests. Therefore, the main outcome measures utilized in this study are the proportions of households in a community with forest access for three types of extraction: fuelwood, fodder and grazing. The dependent variables are thus expressed as access rates. In the sampled communities, forest extraction occurs primarily for household consumption and subsistence production.7 Table 2 presents the access rate by extraction type under different management regimes. A discussion of these results is presented in Section 3.2.

3.2. Forest Management Regimes Forest management regimes affect the conditions of interaction with natural environments, i.e., “who has what kind of access to which kind of natural resources and what use they can make of these resources” (Meynen and Dornboos, 2005). State property regimes have often been viewed as inefficient, ineffectual and especially in the Indian case have been associated with highly restrictive access (Gadgil and Guha, 1995; Guha, 1990). Sociopolitical realities in Himachal Pradesh, however, have forced the government to weaken its control over state forests and hence disallow the indiscriminate exclusion of all forest dwellers (Rangan, 1997). Nevertheless, forest policies continue to be state controlled, “top-down”, bureaucratic and restrictive. About 21% of the sampled communities operate solely under a state management regime. Despite the determination of government to maintain control over state forests, forest dwellers have often invoked the notion of a moral economy to maintain forest rights and to implement traditional management regimes. Such regimes enjoy higher social legitimacy and control over forest access. There is considerable diversity of such traditional institutions in Himachal Pradesh, including scared groves, informal rules that define access and resource management, or general social organizations that include forest management in their scope (Vasan, 2006); some of these regimes have been in existence for decades.8 These regimes operate without aid or recognition from the Forest Department and sometimes despite the efforts of the government agency. About 43.39% of the total sample consists of traditional community institutions. In some respect, these forest management regimes may be superior at achieving the goals of forest management because of their integration into their community, knowledge of existing sociocultural practices, and knowledge of local ecological and social needs. However, they are also subject to limitations with respect to access to forest benefits owing to the perpetuation of existing social inequities through the institutional regime (Adhikari and Di Falco, 2009; Larson et al., 2010; Naidu, 2009; Vasan, 2006).

7 Households occasionally engage in petty trade of forest products but these transactions were unsystematic in nature. 8 The Kangra Forest Cooperatives, for instance, were legalized in 1930s after significant pressure from communities but are not the first known instances of community forest management in Himachal Pradesh. Traditional management regimes, however, typically do not enjoy legal status.

205

Table 2 Forest access proportions by management regime.

Grazing Fodder Fuel

JFM

Community management

State control

Marginals

0.385 0.600 0.673

0.546 0.606 0.792

0.149 0.142 0.928

0.400 0.473 0.806

Note: The p-value of the orthodox test of the hypothesis that forest access proportions are equal across institutional arrangements is approximately zero. However, the p-value is 0.399 for the hypothesis that the proportions utilizing the forest for fodder extraction are equal in the case of JFM and community management.

In addition to traditional regimes, a shift in global and national discourse wrought by certain failures of state regimes and vociferous demands by forest dwellers for increased forest rights led to the adoption of the joint forest management (JFM) program in India. Under JFM, forests are to be co-managed by Forest Departments and communities living in or adjacent to state forests. Proving to be a popular policy tool, the Forest Department, often in conjunction with international aid agencies, has initiated or aided the formation of forest committees to regulate use and management of forests. Himachal Pradesh first initiated forest co-management in a limited area under a JFM scheme9 in 1993 with funding from Department for International Development, a UK-based development agency. Thereafter, additional state-funded community forestry schemes were announced extending the scope of co-management to a wider area within Himachal Pradesh (Vasan, 2003). Under the tenets of co-management, rights and responsibilities concerning forests are vested in the village community, and technical and financial assistance are provided by the state through the Forest Department (Vasan, 2003, 2006). In addition, the comanagement schemes are integrated with development programs to promote the overall wellbeing of the communities (GoHP, 2001). To fulfill objectives of equity, membership is extended to all households in the village (specifically one male and one female member of the household). In Himachal Pradesh, an estimated 500 community-forest institutions were established under JFM by early 2000s (Gouri et al., 2004). Approximately 36% of the sampled communities operate under JFM. While co-management has been touted to be a significant change in forest policy, it has also come under considerable criticism for not being responsive to the needs of local communities and implementing a weak form of democratic participation (Agarwal, 2001; Sarin, 2001; Sundar, 2000; also see Ribot and Larson, 2005; Ribot and Peluso, 2003). In particular, 61% of JFM committees in the sample have been initiated by the local community10 and thus a significant proportion of JFM communities play an active role in the management regime. Moreover, 73.7% of comanagement institutions are registered and hence these communities enjoy some degree of tenure security compared to the case of traditional community and state control regimes. Management regimes regulate what can be extracted, who can extract and how much can be extracted (Agrawal, 1994; Meynen and Dornboos, 2005). For instance, grazing and fodder collection is disallowed during the monsoon periods to allow regeneration of undergrowth in forests. Some communities specify the number of days for which grass can be cut from forests, number of individuals per household allowed to harvest during those days and the quantity of grass that may be extracted. 11 Some communities shut down a portion or the whole forest for a few years to allow for regeneration. Shutting down the forest usually imposes a burden on the members of

9 The subject of forest management in India falls under the “joint” list. Therefore, the central government issues directives but state governments are responsible for implementation of directives and programs pertaining to forests. 10 Initiation requires signatures of a minimum of 50% of voters of the village (gram panchayat ward) (GoHP, 2001). 11 No fees were charged for entering the forest.

206

S.C. Naidu / Ecological Economics 71 (2011) 202–210

Table 3 Univariate statistics of explanatory variables.

Table 4 Quasi-maximum likelihood estimates of fractional logit models for access proportions.

Variable

Mean of proportions

Std dev

Min

Max

prop_lcaste prop_ poor prop_landpoor prop_landrich prop_landcash

0.16 0.21 0.45 0.11 0.17

0.29 0.26 0.34 0.16 0.24

0 0 0 0 0

1 0.97 1 0.76 0.80

Frequency for dummy variables

Percent

shop_d jfm_d traditional_d

70.9 34.55 41.82

Intercept prop_lcaste prop_ poor prop_landpoor prop_landrich prop_landcash

communities unless alternative forest or grazing land is available or households can substitute collection from private lands. Note that both traditional and co-management regimes have been criticized for being insensitive to the needs of women in communities (see Agarwal, 2001, 2007, 2010); during group interviews female members lamented about the increased work burden due to rules imposed by traditional and JFM regimes. However, this aspect was not systematically investigated in this paper. Table 2 compares forest utilization under the three management regimes of interest. A standard chi-squared test reveals that the proportions across regimes are statistically different. Co-management and traditional forest management regimes both display higher access rates (proportions) in grazing and fodder compared to the state management regime. In order to elicit participation, a co-management regime is expected to provide incentives to local communities; this explains higher forest access compared to a management regime under pure state control. A co-management regime needs to provide incentives of higher forest access in order to elicit participation in forest management from local communities so this explains higher access rates under a JFM regime. Further, a traditional community management regime, even if not recognized by the Forest Department, has significant social legitimacy and hence can formulate a collective response to restrictions in forest access. Himachal Pradesh has a history of collective resistance to forest access restrictions and hence this result is to be expected (see Chhatre and Saberwal, 2006; Saberwal, 1999; Vasan, 2006). When community forests are under sole state control without the presence of a community institution, however, the state exercises more (even if imperfect) control over forest access. Further details about the institutional set-up are provided in Naidu (2009). 3.3. Wealth The economy of the study area is agropastoral. Land is thus an important private factor of production that is not only indicative of wealth but also confers social status and provides a safety net in times of economic and environmental shocks (see Agarwal, 1994; Bardhan, 2005; Mearns, 1999). In Himachal Pradesh, land is important but the steep and mountainous terrain necessitates relatively small landholdings. About 80% of cultivators in the state hold less than 2.45 acres (one hectare) of land (GoI, 2001). Thus, landholdings are classified into three categories: less than one acre, between 1 and 2.5 acres and greater than 2.5 acres.12 The average proportion of households in the sampled communities that hold land less than one acre is 45%, 44% hold land between one and 2.5 acres and 11% have landholdings of greater than 2.5 acres. Empirical studies have found that poor households derive a larger proportion of their income from forests and other natural resources compared to wealthier households (Cavendish, 2000; Jodha, 1986; Kamanga et al., 2009). Some others have found results to the contrary (Adhikari, 12 Landlessness is not a common feature in the study area due to a land distribution program carried out in the state in the late 1960s through the 1970s. Nevertheless, differences in landholdings persist.

shop_d shop_rich traditional_d jfm_d jfm_landpoor Observations Robust Wald Chi squared SER Pseudo R2

Grazing

Fodder

Fuel

− 0.815 (0.728) 0.285 (1.313) 3.696*** (1.326) − 1.774* (1.04) − 1.198 (1.694) 4.519*** (1.262) − 0.990 (0.623) 7.524** (3.152) 1.329 (0.859) − 0.163 (0.939) 3.418** (1.529) 52 41.31

− 0.694 (1.067) 1.202 (1.465) 4.429*** (1.567) − 3.138** (1.32) 10.776*** (3.02) 5.288*** (1.525) 0.303 (0.724) − 11.681*** (3.577) 1.112 (0.989) − 1.363 (1.623) 5.549*** (1.734) 52 45.16

3.939*** (1.337) − 0.294 (1.011) 9.194** (4.026) − 2.382 (1.764) 10.244 (6.478) 6.414* (3.461) − 1.297 (1.363) 2.034 (17.529) − 1.488* (0.812) − 5.817** (2.307) 4.875 (3.008) 53 26.16

0.573 0.363

0.547 0.385

0.322 0.563

Note 1: While data were collected from 56 forest communities, missing data necessitated discarding some observations. Note 2: Estimates are expressed on the logit scale. The quantities in (.) are estimated robust standard errors. SER denotes the estimated standard error of the regression. Wald statistics test the joint significance of the estimated coefficients. Significance codes are defined as ‘***’ for the 0.01 level, ‘**’ for the 0.05 level and ‘*’ for the 0.1 level.

2005; Adhikari B. et al., 2004; Cavendish, 2000; Narain et al., 2008b; Vedeld et al., 2007) and still others have provided evidence in favor of a non-linear relationship between wealth, income and forest use (Narain et al., 2008a; Reddy and Chakravarthy, 1999). Land might be an endogenous variable if greater access to forests leads to higher wealth, i.e., landholdings. However, land markets in rural India, including the study area, are extremely thin (Bardhan, 2005), and conversations with state forest officials confirm that encroachments on forestland are not a serious issue. Accordingly, land variables are taken to be exogenous in the econometric model. Since land connotes economic and political power, those with large holdings of land are expected to benefit relatively more from forest extraction. However, the net effect of land will depend on whether the forest products are complementary inputs in the rural production process. Further, in the econometric model an interaction term between land-rich households and the JFM regime is introduced to ascertain the effect of the co-management policy adopted by the government. Additionally, private land may be correlated with poverty but it does not fully represent the extent of consumption poverty. Therefore, the variable prop_ poor measures the proportion of households unable to supply enough food for the entire year using either household food production or earned income. Mamo et al. (2007) employ a similar measure of food deficiency. This variable (with a mean of 0.21)13 has a statistically significant positive correlation (of 0.37) with households in the prop_ landpoor category and a low negative statistically insignificant relationship (of −0.15) with households in the prop_ landrich category. This variable could also be considered endogenous if higher access leads to lower poverty. The variable was tested for endogeneity as described by Papke and Wooldridge (2008) and Wooldridge (2005) but proved not to be endogenous at the conventional levels of significance. The

13 Note that the mean for the sample is congruent with the official poverty rate for the districts Kangra (0.219) and Mandi (0.20) (GoHP, 2007).

S.C. Naidu / Ecological Economics 71 (2011) 202–210

variable is expected to positively affect the dependent variables due to the importance of forests in the rural economy. Brief explanations for other variables are provided in Table 1. 4. Results and Discussion In the regression model, three alternative measures of forest access are employed. The dependent variables measure the proportion of total households in a community that extract fodder, graze their livestock in community forests, and collect fuelwood. The data are prone to clustering the endpoints and hence the conditional expectation function of these data is nonlinear. An OLS application hence is unsuitable. In contrast, the family of generalized linear models relaxes a number of crucial assumptions and enables the analyst to implement a distribution appropriate to the problem under consideration. Moreover, generalized linear models are an extension of classical linear models such that the probability density function of the response variables belongs to the exponential family (of which the normal distribution is a special case). This paper estimates a generalized linear model under the maintained hypotheses of binomial distribution and logit link function, which is also described as a fractional logit model in the econometrics literature.14 Quasi-maximum likelihood estimates (QMLE) of the model parameters are reported. Note that the QMLE is consistent under certain forms of misspecification. (Cox and Snell, 1989; McCullagh and Nelder, 1983; Papke and Wooldridge, 1996). Table 3 presents descriptive statistics for the explanatory variables. The regression results are reported in Table 4, along with the robust standard errors as suggested by Papke and Wooldridge (1996) and Wooldridge (2002) .15 4.1. Forest Management Regimes The coefficients of the dummies for traditional management (traditional_d) and co-management (jfm_d) regimes are statistically significant and negative for the fuelwood regression presented in Table 4. This suggests that community regimes, whether traditional or JFM regimes decrease access rate to benefits from firewood collection. This finding is consistent with the analysis of Table 2. Since the access rate does not measure the absolute or relative level of extraction, one cannot infer that lower access necessarily results in lower absolute levels of extraction or higher levels of conservation. Instead, one can conclude that a lower proportion of households is able to benefit from firewood extraction when one of these two management regimes is in operation. Fuelwood is an important component in household production and reproduction, the collection of which is carried out by women in Himachal Pradesh (Negi, Rana and Sharma, 1997 cited in GoHP, 2002) and many other parts of India. It is a major input to the economic activities engaged by women. A major criticism against various forms of community forestry, whether traditional management regimes or more structured forms of participatory or co-management regimes, is lack of attention to the impact of restrictions. The ability to access forest for fuelwood is impacted by closure of forests for regeneration or limits placed on extraction. The statistically insignificant coefficient of the fair price shop (shop_d) suggests that alternatives to fuelwood may

14 Although the beta distribution is an alternative to the binomial under certain conditions, the former distributional assumption is not plausible when a significant number of observations occur at unity or zero as in this sample. 15 In an alternative specification, an interaction term between land-poor households and traditional community regime was substituted for the interaction term between land-poor households and co-management regime dummy. Both models perform well and their results are comparable. Other environmental variables and a district dummy were included as controls in both specifications but because they were consistently statistically insignificant they were excluded to conserve degrees of freedom. Various diagnostic checks were conducted, including inspection of VIFs, distributional properties of the deviance residuals, and plots of residuals against covariates and the linear predictor. On balance, the three models are robust and provide good fits for these data.

207

not be available or the price may be prohibitive for most users. It appears that compared to the baseline case of state regimes, JFM and traditional management regimes place a high cost on fuelwood collectors, i.e., women. It is likely that households and particularly women have to travel further distances to acquire fuelwood from other forests, engage in illegal extraction, or utilize agricultural wastes that serve as inferior substitute fuels (Agarwal, 2001, 2010; Agrawal and Gibson, 2001),16 all of which induce significant psychological and health costs. The dummy variables for co-management regimes are statistically insignificant in determining access rates for fodder and grazing at the 10% level of significance. Historically, the Forest Department in Himachal Pradesh has adopted an alarmist discourse of environmental collapse to impose severe restrictions on grazing. Saberwal (1999) in his study of pastoralism in Himachal Pradesh attributes the discourse of the Forest Department to the struggle over control over forestlands; the Forest Department has exerted significant effort to restrict grazing of sheep and goats. Thus, in a co-management regime, the presence of the Forest Department is unlikely to create a distinction between access rates vis-à-vis a regime under complete state control. With respect to traditional management regimes, however, Table 2 indicates that this regime allow the highest rate of access for grazing across the three management regimes. In the econometric analysis, the positive coefficient for traditional management regimes is statistically significant only at a 12% level, which indicates weak support for increased access under a traditional management regime relative to a state management regime. Further, the interaction term between land-poor and co-management regimes is positive and highly statistically significant. In this sample, the proportion of grazing households and fodder collectors increases for land-poor households conditional on the presence of a co-management regime. The various co-management schemes undertaken by the Forest Department under the direction of the state government combines forest conservation with development work. For instance, the state government order on Sanjhi Van Yojana, one of the co-management schemes implemented since early 2000s, cites as an objective the creation of physical and social infrastructure for use by communities for poverty reduction and linking with food for work programs for welfare enhancement (GoHP, 2001). Additionally, the government of Himachal Pradesh has been incentivizing adoption of hybrid cows instead of small ruminants and creating markets for milk products. These measures are likely to improve the expected or perceived return on livestock holdings thus increasing access rates for both forest activities. While this may be incongruous with the discussion on the aversion of the Forest Department to grazing, it should be noted that land-poor households are likely to hold fewer livestock compared to relatively land-rich households, which can be easily regulated and monitored.

4.2. Wealth The positive sign on the variable prop_poor suggest that there is a positive correlation between access rate and proportion of food deficient households in the community. However, the coefficient for the variable proportion of land-poor households is negative and statistically significant for the fodder and grazing equations. If land is a true measure of wealth, land-poor households are less likely to hold livestock because of the initial investment required. The inability of less wealthy households to benefit from this livelihood activity, which not only yields market income through the sale of animal products but also contributes to household consumption and may be used in the event of economic shocks, means that the derived demand for forests is lower. Land-poor households are less likely to access forests to graze animals or collect fodder, all else constant.

16 Fuelwood is collected only for household consumption and production and not for its commercial value.

208

S.C. Naidu / Ecological Economics 71 (2011) 202–210

The coefficient for the variable prop_landrich is positive and highly statistically significant for the fodder equation. A higher proportion of landrich households in the community increases the rate of forest access. This suggests that land rich households are more likely to participate in fodder extraction than households in other land categories. However, the variable is statistically insignificant in the grazing equation. To make sense of the differing impact of higher landholdings, one has to consider the context in which livestock rearing is carried out in the study region. Wealthier households not only have the economic capacity to purchase and maintain livestock but also are more likely to purchase improved breeds of Jersey cows. This breed of cows yields higher quantities of milk and thus has a higher total market return. However, it is expensive to purchase and according to respondents is a poor grazer; unlike indigenous varieties of domesticated ruminants, they are unsuited to the mountainous terrain. This breed must thus be primarily stall-fed thereby increasing the demand for fodder collection. Mamo et al. (2007) argue with respect to their dataset from Ethiopia that wealthy households are likely to decrease forest extraction due to alternatives, at the same time because they have large livestock holdings, they are more likely to engage in fodder extraction (also see Heltberg, 2001). However, the data suggest that the net effect on the access rate for fodder collection when rich households have easy access to a shop is negative as the proportion of land-rich households increases, as suggested by the negative sign on the interaction term shop_ rich. 5. Conclusions The vast majority of the rural households in the global south live in and around forests, 350 million people in Asia (Chomitz et al., 2007 in Mahanty et al., 2009), and 400 million in Africa (CIFOR, 2005 in Coulibaly-Lingani et al., 2009). Irrespective of the actual number, it is widely accepted that forests constitute an important economic resource in rural livelihoods. Even though about 80% of the world's forests are owned by governments (FAO, 2010), public ownership does not guarantee forest access to rural populations. Many forest policies around the world, especially in the developing world, suffer from a colonial hangover whereby humans are considered inimical to the objective of conservation. Increasing attempts to secure forest rights through community forestry initiatives have lessened the fallout of restrictions. Nevertheless, rights do not imply access (Larson et al., 2010; Rangan, 1997; Ribot and Peluso, 2003). At the time of fieldwork, Himachal Pradesh was one of the few states in India wherein forest dwellers claimed traditional usufruct rights. However, the passage of the Forest Rights Act in 200617 legalized the rights to forests for bonafide personal consumption across the country. In addition, the Act also endows forest communities with the right to manage their community forests; this constitutes a threat to state control and forms of co-management regimes that are less democratic.18 Yet, there has been no study that compares forest access under different existing management regimes. In addressing this lacuna in the literature, this paper studies the influence of private wealth and management regimes on forest access. Forest access was measured in terms of proportion of households that utilize the forest for fodder, fuelwood collection, and grazing livestock. In other words, rather than measure the aggregate level of extraction of total benefits from forests, forest access is represented by a proportion. The study area proved to be amenable to a comparison of forest access across communities for two reasons. First, all households in the community have similar forest rights so the issue of usufruct rights could be controlled. Second, these usufruct

17 The Forest Rights Act, 2006 has been hailed as an important legislation that has the potential to reverse historical injustices by providing tenure security thereby increasing rural wellbeing and forest health. 18 The continuance of JFM under the Forest Rights Act 2006 is highly contested.

rights are vested in the community, thereby making it possible to adopt the community as the unit of analysis. The results of the statistical analysis offer insights into the effects of management regimes and wealth, and suggests the following. First, forest access rates for fodder and grazing are lowered as proportion of landpoor households increases, implying that the less wealthy are unable to benefit from forest access, all else constant. On the other hand, as the proportion of land-rich households increases the access rate for fodder collection in a community rises. Second, management regimes play a significant role in determining access. Access to grazing may be higher under a traditional community regime but access rates for fuelwood are lower under traditional community as well as co-management regimes compared to the state control regime. While this could be a result of access restrictions under the former two regimes, it is possible that members of forest communities are more likely to adhere to restrictions decided by the community as opposed to those imposed by pure state control. Third, the interaction terms between co-management regimes and proportion of land-poor households suggests that this regime has the potential to provide forest access to the least wealthy in forest communities. The last result appears to be contrary to the literature pertaining to co-management in South Asia. Co-management in South Asia has been criticized for inequitable outcomes, the retention of government control over forest and subversion of democratic processes while at the same time obtaining free labor to monitor and enforce government determine forest rules and policies (e.g., Sarin, 2001; Sundar, 2000) and perpetuating existing inequities through the distribution of benefits (e.g., Adhikari, 2005; Agarwal, 2001; Iversen et al., 2006; Thoms, 2008). However, in understanding higher access rates for land-poor households under a co-management relative to a state management regime, the following factors should be considered. First, approximately 61% of co-management institutions have been set up at the initiative of the local community. This fact indicates that the presence of active members and leaders within the community may negotiate better terms of forest rights for community members. This is relevant in the study area because rights are vested in the community and not individuals or households. However, an increase in access rates (or the ability to benefit from forest rights) does not indicate whether land-poor households are able to benefit more in absolute or relative terms.19 A second related point pertains to the history of forest use and management in the region. Past attempts by the state government to restrict forest use in different regions in Himachal Pradesh (see for instance, Saberwal, 1999; Chhatre and Saberwal, 2006; Vasan, 2006) have been met with stiff resistance. Thus, under active community participation, it would be difficult to restrict human presence in forests. Besides, unless the community perceives net benefits from participation, co-management regimes are likely to be failures. Third, one cannot discount the role of tenure security available under co-management but unavailable under state control and traditional management regimes. In short, this paper does not offer evidence against the valid criticisms against JFM regimes in India. Rather it suggests that access rates are higher in forests under co-management regimes, not necessarily borne out of state benevolence, but likely due to an active community. While this paper offers insights into the complexities associated with forest management regimes, more systematic research is needed to ascertain the processes and mechanisms underlying management regimes. Recent research into the issue of “authority” underlying management regimes is a step in that direction (e.g., Larson et al., 2010; Ribot et al., 2008).

19 Vasan (2003), in a comparison of various state-initiated and traditional community regimes in Himachal Pradesh argues that co-management regimes are more likely to include marginalized groups of individuals in the decision-making process and hence lead to equitable outcomes. Validation of this claim would require data on absolute and relative levels of access. I thank an anonymous reviewer for pointing out the difference between equitable and equal access.

S.C. Naidu / Ecological Economics 71 (2011) 202–210

Acknowledgements Comments were provided by Thomas Stevens, James Boyce, Sylvia Brandt and the UMASS Environmental Working Group on earlier versions of this paper. The paper benefited significantly from comments and suggestions by Panayiotis Manolakos and three anonymous reviewers. I am grateful to Sandeep Minhas who helped in data collection, and the residents of Kangra and Mandi districts in Himachal Pradesh who participated in interviews. The Department of Resource Economics, University of Massachusetts Amherst provided financial support for fieldwork. The usual disclaimer applies.

References Ackerman, F., 2008. Poisoned for pennies: the economics of toxics and precaution. Island Press, Washington DC. Adhikari, B., 2005. Poverty, property rights and collective action: understanding the distributive aspects of common property resource management. Environment and Development Economics 10, 7–31. Adhikari, B., Di Falco, S., 2009. Social inequality, local leadership and collective action: an empirical study of forest commons. European Journal of Development Research 21, 179–194. Adhikari, B., Di Falco, S., Lovett, J.C., 2004. Household characteristics and forest dependency: evidence from common property forest management in Nepal. Ecological Economics 48, 245–257. Adhikari, M., Nagata, S., Adhikari, M., 2004. Rural household and forest: an evaluation of households dependency on community forest in Nepal. Journal of Forest Research 9, 33–44. Agarwal, B., 1994. Gender and command over property: a critical gap in economic analysis and policy in South Asia. World Development 22, 1455–1478. Agarwal, B., 2001. Participatory exclusions, community forestry, and gender: an analysis for South Asia and a conceptual framework. World Development 29, 1623–1647. Agarwal, B., 2007. Gender inequality, cooperation and environmental sustainability. In: Baland, J.M., Bardhan, P., Bowles, S. (Eds.), Inequality and Environmental Sustainability. Princeton University Press, Princeton, NJ, pp. 274–313. Agarwal, B., 2010. Gender and Green governance: the political economy of women's presence within and beyond community forest. Oxford University Press, New York. Agrawal, A., 1994. Rules, rule making, and rule breaking: examining the fit between rule system and resource use. In: Ostrom, E., Gardner, R., Walker, J. (Eds.), Rules, Games, and Common-Pool Resources. University of Michigan Press, Ann Arbor, MI, pp. 267–282. Agrawal, A., Gibson, C.C., 2001. Communities and the environment: ethnicity, gender, and the state in community-based conservation. Rutgers University Press, Piscataway, NJ. Bardhan, P., 2005. Scarcity, conflicts, and cooperation: essays in the political and institutional economics of development. MIT Press, London. Beck, T., 1994. Common property resource access by the poor and class conflict in West Bengal. Economic and Political Weekly, pp. 187–197. Beck, T., Ghosh, M.G., 2000. Common property resources and the poor: findings from West Bengal. Economic and Political Weekly 35, 147–153. Beck, T., Nesmith, C., 2001. Building on poor people's capacities: the case of common property resources in India and West Africa. World Development 29, 119–133. Bromley, D. (Ed.), 1992. Making the commons work: theory, practice and policy. ICS Publications, San Francisco. Burkett, P., 2006. Marxism and Ecological Economics: Toward a Red and Green Political Economy. Brill, Boston. Cavendish, W., 2000. Empirical regularities in the poverty–environment relationship of rural households: evidence from Zimbabwe. World Development 28, 1979–2003. Chhatre, A., 2003. The mirage of permanent boundaries: politics of forest reservation in the Western Himalayas, 1875–97. Conservation and Society 1, 137–159. Chhatre, A., Saberwal, V.K., 2006. Democratizing nature: politics, conservation, and development in India. Oxford University Press, New Delhi. Coomes, O.T., Barham, B.L., Takasaki, Y., 2004. Targeting conservation-development initiatives in tropical forests: insights from analyses of rain forest use and economic reliance among Amazonian peasants. Ecological Economics 51, 47–64. Coulibaly-Lingani, P., Tigabu, M., Savadogo, P., Oden, P., Ouadba, J., 2009. Determinants of access to forest products in southern Burkina Faso. Forest Policy and Economics 11, 516–524. Cox, D.R., Snell, E.J., 1989. Analysis of Binary Data. Chapman and Hall, New York. Cranney, B., 2001. Local environment and lived experience: the mountain women of Himachal Pradesh. Sage Publications, Thousand Oaks, CA. FAO, 2010. Global Forest Resources Assessment 2010. Food and Agriculture Organization of the United Nations, Rome. Forest Survey of India, 2009. India: State of Forest Report 2009. Ministry of Environment and Forestry, Government of India, Dehradun, India. Gadgil, M., Guha, R., 1995. Ecology and equity: the use and abuse of nature in contemporary India. Routledge, London. GoHP, 2001. Sanjhi Van Yojana. Notification File No FFE-C (9).1/2001. http://hpforest. nic.in/for_mgt.htm 2001. GoHP, 2002. Himachal Pradesh human development. Government of Himachal Pradesh, Shimla.

209

GoHP, 2007. Socio-Economic Indicators of Himachal Pradesh. Department of Economics and Statistics, Government of Himachal Pradesh, Shimla. GoI, 2001. Census of India Series 1: Provisional Population Totals Paper 1. Office of the Registrar General & Census Commissioner, India, New Delhi. Gouri, Mudgal, S., Morrison, E., Mayers, J., 2004. Policy influences on forest-based livelihoods in Himachal Pradesh, India Policy Analysis Process Paper. Stockholm Environment Institute. Guha, R., 1990. The Unquiet woods: ecological change and peasant resistance in the Himalaya. University of California Press, Berkeley. Heltberg, R., 2001. Determinants and impact of local institutions for common resource management. Environment and Development Economics 6, 183–208. Iversen, V., Chhetry, B., Francis, P., Gurung, M., Kafle, G., Pain, A., Seeley, J., 2006. High value forests, hidden economies and elite capture: evidence from forest user groups in Nepal's Terai. Ecological Economics 58, 93–107. Jodha, N.S., 1986. Common property resources and the rural poor in dry regions of India. Economic and Political Weekly 21, 1169–1182. Jodha, N.S., 1995. Common property resources and the environmental context: role of biophysical versus social stresses. Economic and Political Weekly 30, 3278–3283. Kamanga, P., Vedeld, P., Sjaastad, E., 2009. Forest incomes and rural livelihoods in Chiradzulu District, Malawi. Ecological Economics 68, 613–624. Khan, S.R., Khan, S.R., 2009. Assessing poverty–deforestation links: evidence from Swat, Pakistan. Ecological Economics 68, 2607–2618. Larson, A.M., 2008. Land tenure rights and limits to forest management in Nicaragua's North Atlantic Autonomous Regime: Making the rules of the game. Paper presented at IASC, July 14–18, Cheltenham, England. Larson, A.M., Barry, D., Dahal, G.R., Colfer, C.J.P. (Eds.), 2010. Forests for People: Community Rights and Forest Tenure Reform. Earthscan, London. Mahanty, S., Guernier, J., Yasmi, Y., 2009. A fair share? Sharing the benefits and costs of collaborative forest management. International Forestry Review 11, 268–280. Mamo, G., Sjaastad, E., Vedeld, P., 2007. Economic dependence on forest resources: a case from Dendi District, Ethiopia. Forest Policy and Economics 9, 916–927. McCullagh, P., Nelder, J.A., 1983. Generalized Linear Models. Chapman and Hall, New York. Mearns, R., 1999. Access to Land in Rural India - Policy And Options. World Bank Policy Research Working Paper No. 2123. Washington, DC. Meynen, N., Dornboos, M., 2005. Decentralising natural resource management: a recipe for sustainability and equity? In: Ribot, J.C., Larson, A.M. (Eds.), Democratic Decentralisation through a Natural Resource Lens. Routledge, New York. Morrison, E., 2001. Participatory forestry in Himachal Pradesh India: Improving policylivelihood relationships in South Asia. Stockholm Environment Institute: Policy Review Paper No. 2. Naidu, S.C., 2009. Heterogeneity and collective management: evidence from common forests in Himachal Pradesh, India. World Development 37, 676–686. Narain, U., Gupta, S., Van 't Veld, K., 2008a. Poverty and the environment: Exploring the relationship between household incomes, private assets, and natural assets. Land Economics 84, 148–167. Narain, U., Gupta, S., van 't Veld, K., 2008b. Poverty and resource dependence in rural India. Ecological Economics 66, 161–176. Ostrom, E., 1990. Governing the commons: The evolution of institutions for collective action. Cambridge University Press, New York. Papke, L.E., Wooldridge, J.M., 1996. Econometric methods for fractional response variables with an application to 401(k) plan participation rates. Journal of Applied Econometrics 11, 619–632. Papke, L.E., Wooldridge, J.M., 2008. Panel data methods for fractional response variables with an application to test pass rates. Journal of Econometrics 145, 121–133. PROBE Team, 1999. Public report on basic education in India. Oxford University Press, New Delhi. Rangan, H., 1997. Property vs. control: The state and forest management in the Indian Himalaya. Development and Change 28, 71–94. Reddy, S.R.C., Chakravarthy, S.P., 1999. Forest dependence and income distribution in subsistence economy: evidence from India. World Development 27, 1141–1149. Ribot, J.C., Larson, A.M. (Eds.), 2005. Democratic decentralisation through a natural resource lens. Routledge, New York. Ribot, J., Peluso, N., 2003. A theory of access. Rural Sociology 68, 153–181. Ribot, J.C., Chhatre, A., Lankina, T., 2008. Introduction: institutional choice and recognition in the formation and consolidation of local democracy. Conservation and Society 6, 1–11. Saberwal, V.K., 1999. Pastoral politics: shepherds, bureaucrats, and conservation in the western Himalaya. Oxford University Press, New York. Sarin, M., 2001. Disempowerment in the name of participatory forestry? Village joint forest management in Uttarakhand. Forests, Trees, and People Newsletter No. 44. Schlager, E., Ostrom, E., 1992. Property-rights regimes and natural resources: a conceptual analysis. Land Economics 68, 249–262. Sundar, N., 2000. Unpacking the ‘joint’ in joint forest management. Development and Change 31, 255–279. Thoms, C.A., 2008. Community control of resources and the challenge of improving local livelihoods: a critical examination of community forestry in Nepal. Geoforum 39, 1452–1465. Vasan, S., 2003. Community forestry: historical legacy of Himachal Pradesh. Himalayan Research Bulletin 21, 36–45. Vasan, S., 2006. Living with diversity: forestry institutions in the western Himalaya. Indian Institute of Advanced Study, Shimla. Vedeld, P., Angelsen, A., Sjaastad, E., Berg, G.K., 2004. Counting on the environment: Forest incomes and the rural poor. Environmental Economics Series. Paper No. 98. World Bank, Washington D.C. Vedeld, P., Angelsen, A., Bojö, J., Sjaastad, E., Kobugabe, Berg G., 2007. Forest environmental incomes and the rural poor. Forest Policy and Economics 9, 869–879. Vlachou, A., 2002. Nature and value theory. Science and Society 66, 169–201.

210

S.C. Naidu / Ecological Economics 71 (2011) 202–210

Wade, R., 1988. Village republics: economic conditions for collective action in South India. Cambridge University Press, New York. Wilshusen, P.R., 2009. Shades of social capital: elite persistence and the everyday politics of community forestry in southeastern Mexico. Environment and Planning A 41, 389–406. Wooldridge, J.M., 2002. Econometric analysis of cross section and panel data. MIT Press, Cambridge, Mass.

Wooldridge, J.M., 2005. Unobserved heterogeneity and estimation of average partial effects. In: Andrews, D.W.K., Stock, J.H. (Eds.), Identification and Inference for Econometric Models: Essays in Honor of Thomas Rothenberg. Cambridge University Press, Cambridge, pp. 27–55. World Bank, 2005. Unlocking opportunities for forest-dependent people in India. World Bank, Washington, D.C.